Adding framework for compartmental models

[WillemWybo]

The present PR implements compartmental models (cm) for NEST. It adds a family of files, prefixed by 'cm_' to the models/ directory. These files implement the cm.

Furthermore, the PR adds tests for the cm (pynest/nest/tests/test_compartmental_model.py), as well as examples (pynest/examples/compartmental_model).

The dendritic structure (branching, parameters,...) is determined by the user at runtime. For this, the PR adds a function nest.AddCompartment() to the python API, as well as a function nest.AddReceptor(), which allows the user to add one of three receptor types ('AMPA', 'GABA' or 'AMPA+NMDA'). The number of receptors to a compartment is entirely user-defined.

Currently, both the ion-channels and receptors types are hard-coded into the model. We are working to include user-specified ion-channels and receptor types through NESTML. This NESTML functionality will then render the specific versions of models/cm_etype.* and models/cm_syns.* implemented here obsolete.

[jougs]

@WillemWybo, please fix the formatting errors either manually or by installing clang-format-9 and running build_support/format_all_c_c++_files.sh

[jougs]

Many thanks for this wonderful addition and the endurance during the review process ;-)

Just one additional request from my side: Can you please add a paragraph about the new models (possibly with a short example?) to the release notes for NEST 3.3, so users can discover them more easily?

[WillemWybo]

Many thanks for this wonderful addition and the endurance during the review process ;-)

Just one additional request from my side: Can you please add a paragraph about the new models (possibly with a short example?) to the release notes for NEST 3.3, so users can discover them more easily?

Added it to the release notes!

[pnbabu]

Thanks!

[jougs]

Many thanks for this contribution.